You spent five months on it.
Design, arguments, a rewrite,
the three edge cases
that turned out to be the actual product.
Somebody will reproduce
the visible part of it
in a weekend.
Not because they are better.
Because the hard part
was never the typing,
and you published the answer
to the hard part
on your own marketing page.
Most of what those five months bought
was knowledge that the thing was possible,
that customers wanted it in that shape,
and that the obvious simpler version
does not work.
That knowledge is now free,
and you gave it away
the day you shipped,
as everybody always has.
What is new is the rest.
Implementation used to be a moat
because implementation was slow.
A competent person
with a clear specification
needed months,
and you had months of head start
simply because building is hard.
Building is getting cheap
at a rate nobody has finished
adjusting to,
and a feature is a specification
with a screenshot attached.
So stop defending the feature.
It is not defensible
and it never really was.
Spend your attention
on the things that stay slow
no matter how fast code gets written.
Four years of your customers' data,
which cannot be rebuilt over a weekend
by anyone, at any speed.
The integration nobody wants to do,
with the ancient system
that half your market runs on.
The paperwork.
The certification.
The regulator who took eleven months
to answer a question.
The fact that you have run this
for four years
and never lost anybody's work,
which is a claim
a new thing cannot make
at any price.
The support answer at the weekend
from somebody who understands
what the customer was trying to do.
None of that copies.
And turn it round,
because it cuts both ways.
The incumbent with the feature
you have always envied
is no longer protected by it.
Go and build theirs in a weekend.
Then spend your five months
on whatever they have
that a weekend cannot produce.
– Serguey Asael Shinder
